The research will examine drinkers and drinking at outlets to identify social mechanisms by which drunken driving and alcohol-related violence are maintained in communities. Patrons in 165 bars in six California communities will be surveyed on exit from bars on weekend nights, followed by a phone survey. The exit survey will characterize drinking and social interactions that evening. The telephone survey will provide background measures of drinkers'characteristics and risk related behaviors (N H 1,620). Patron and staff interviews in select outlets identified by risk for drinking-driving and violence, together with bar observations conducted in a random sample of the total bars, will characterize social relations and norms that may reinforce problem behaviors. These qualitative assessments of bar social dynamics will inform and extend quantitative studies. Three hypotheses will be tested: (1) Greater neighborhood social disorganization will be related to greater rates of drunken driving and aggression among bar patrons. (2) Greater bar concentrations will be related to greater clustering of patrons across establishments correlated with risk characteristics. (3) Greater clustering by patrons will result in greater social cohesion in high-risk outlets. The short-term goal of the proposed research is to determine whether at-risk drinkers cluster across on-premise alcohol outlets and whether clustering is increased in areas with high outlet concentrations. The mid-term goal of the study is to initiate studies of social processes that maintain problem behaviors within outlets, identifying social structures that make problems resilient to change. The long-term goal is to provide communities with guidance on regulatory processes that may ameliorate neighborhood problems related to alcohol outlets. PUBLIC HEALTH RELEVANCE: The aim of this study is to understand how drunken driving and alcohol-related violence are maintained in communities, in order to determine whether at-risk drinkers cluster in specific bars within communities, and whether clustering is increased in areas with many bars. We are proposing to conduct a telephone survey of bar patrons, recruited through a doorway survey, from 165 bars in six California communities, together with interviews with patrons and staff in a subset of bars identified as more likely to encourage drunken driving and violence, and observations in a random sample of all bars. The study aims to test these three hypotheses: (1) Higher rates of neighborhood "social disorganization" (based on relative socio-economic status of residents) will be related to higher rates of drunken driving and aggression among bar patrons;(2) More bars in an area will be related to more clustering, within these bars, of patrons according to their likelihood for problematic drinking, drunken driving, and violence;and (3) More clustering by patrons in this way will result in tighter social relations among people in bars identified with drunken driving and violence.